TrojAI provides an enterprise AI security platform designed to secure AI models, applications, and agents across the development and production lifecycle. Its positioning centers on helping organizations identify risky model behavior before deployment and enforce protections against active threats at run time, with support for complex enterprise environments and self-hosted deployments.
TrojAI’s platform is built for organizations that need AI security controls to keep pace with expanding attack surfaces across generative AI, copilots, and agentic workflows. The company emphasizes customizable policy controls, scalable production performance, and security standard mapping to help enterprises operationalize AI more safely.
Offerings, Capabilities, and Integrations
TrojAI’s capabilities span build-time assessment and run-time enforcement. It supports automated AI red teaming for commercial, open-source, and custom models, and it provides real-time inspection of prompts, responses, tools, and agent traffic to address risks such as prompt injection, jailbreaks, sensitive data disclosure, toxic content, system prompt leakage, and other unsafe behaviors. It also supports custom policies and maps findings to frameworks such as OWASP, MITRE, and NIST.
TrojAI integrates into enterprise environments through self-hosted deployment options, browser extensions for governing employee use of third-party GenAI tools, and connections into SIEM, SOAR, ticketing, and audit workflows. The company has also published integrations with OpenAI’s ChatGPT Enterprise Compliance API, TrueFoundry AI Gateway, Wiz AI-APP, and JFrog Artifactory and Evidence to extend AI security into compliance, gateway control, asset discovery, and DevSecOps pipelines.
Products and Services
- TrojAI Detect: TrojAI Detect is TrojAI’s build-time AI red teaming offering. It automatically tests AI models, applications, and agents for behavioral risk using static, manipulated, dynamic, multi-turn, and agentic attack methodologies, and it supports tabular, NLP, and LLM workloads with more than 150 built-in tests plus custom testing options.
- TrojAI Defend: TrojAI Defend is a run-time AI firewall that monitors and secures AI models, applications, and agents in production. It can monitor, alert, block, redact, and log traffic, applies customizable policy rules to inputs and outputs, offers browser extensions for third-party GenAI tools, and supports integrations into SIEM, SOAR, and ticketing systems.
- TrojAI Defend for MCP: TrojAI Defend for MCP extends run-time protection to Model Context Protocol deployments and agentic AI workflows. It focuses on discovery and control of MCP servers and tools, blocking unapproved paths, detecting tool or server tampering, enforcing MCP-specific policies, and producing audit logs for governance and incident response.
Target Customers
TrojAI targets large enterprises building or operating AI models, applications, copilots, and agents in production. Its messaging is oriented toward organizations with demanding security, governance, and scale requirements, including Fortune 500 environments and enterprises managing complex AI deployments across multiple models and clouds.
Its platform is relevant to security teams, CISOs, AI/ML teams, platform engineering groups, DevSecOps organizations, and compliance stakeholders that need visibility into AI behavior before and after deployment. TrojAI is also positioned for regulated or policy-sensitive environments, including use cases in finance, healthcare, legal, and government, and for enterprises that need to govern employee use of third-party GenAI applications.
Cloud Integrations and Marketplace
- Microsoft Marketplace: TrojAI has verified Microsoft Marketplace presence, including listings for the broader TrojAI AI Security Platform as well as TrojAI Detect and TrojAI Defend. These listings position the company for Microsoft-aligned cloud procurement and deployment workflows.
Key People
- Lee Weiner: CEO
- James Stewart, PhD: CTO and Co-Founder
- Stephen Goddard: COO and Co-Founder
- Stan Petley: Director of Engineering
- Phil Munz: Senior Manager, Data Science
- Julie Peterson: Product Marketing
- Christian Falco: Partnerships
Key Facts
- Headquarters: Saint John, New Brunswick, Canada
- Employees: Approximately 19 employees
- Annual Revenue: Approximately US$1.9M
- Parent Company: None
- Subsidiaries: None
- Publicly Listed: Privately held
Analyst Recognitions
- Gartner: 2025 Market Guide for AI Trust, Risk and Security Management – Representative Vendor. 2024 Hype Cycle for Emerging Technologies – Representative Vendor. 2024 Hype Cycle for Generative AI – Representative Vendor.
